ETD

• bioreactor
• brain organoids
• electrophysiology
• MEA
• millifluidic system

In this thesis, I developed a novel millifluidic system to enable the electrophysiological characterisation of whole-brain organoids while cultured in dynamic conditions. The work was carried out in the framework of the Horizon EIC project entitled “NAP: twin-on-a-chip brains for monitoring individual sleep habits”, which aims at using brain organoids to study sleep disorders as early predictive signs of Parkinson’s disease.
Preliminary finite element models were employed to identify the flow rate at the inlet of the bioreactor ensuring the best trade-off between oxygen supply and shear stress applied to the surface of the construct. For this purpose, dimensionless numbers – such as the Graetz number and Thiele modulus - were also evaluated to better understand the balance between oxygen transport and consumption mechanisms occurring in the bioreactor.
Based on the insights from finite element analysis, a novel millifluidic system was created: it combines the upper part of a commercial laminar flow bioreactor and a multi-electrode array chip, used as the system bottom where organoids are meant to be grown. Two types of multi-electrode arrays were employed for this purpose, namely planar ones as commercially available and the prototype of a 3D multi-electrode array designed within the NAP project. To properly assemble the two components, a silicone (more specifically, PDMS) adapter ring was designed and fabricated by casting. Two different mould designs were evaluated, so as to optimize the production process for the adapter. The system is actuated by means of a peristaltic pump, applying the inlet flow rate identified through finite element analysis. A locking system was also developed to tighten the millifluidic system, thus ensuring a watertight seal with no leakages over 48 hours – with respect to a typical duration of electrophysiological recordings of a few minutes.
Prior to recording, viability tests were performed to verify that the exposure to the flow rate selected through finite element analysis does not harm cells within the organoid. These tests were based on the Trypan Blue exclusion assay, performed after 48 hours of dynamic culture in the developed system, and showed that dynamic culture conditions do not critically affect cell survival, allowing for reliable electrophysiological characterization of the constructs.
Electrical signals generated by the constructs were recorded under a laminar flow hood by connecting the millifluidic system to a commercial recording system, whose headstage was directly interfaced with the multi-electrode array chips. Four-minute acquisitions were performed in both static and dynamic conditions, with organoids grown on either the planar or the 3D multi-electrode array chip. The outcoming recordings were then processed through an algorithm purposely developed within the NAP project. After an initial filtering step, it detects neural spikes in order to estimate relevant parameters which characterize the functional activity of neurons within the construct. Specifically, the mean firing rate and the burstiness index were considered for the purpose of the thesis. The mean firing rate describes the average rate of action potentials produced by neurons and recorded in a specific time window; the burstiness index is a measure of how much neural activity is organized in groups of close-up spikes rather than in isolated peaks.
Using the 3D multi-electrode array, the recorded signals display high noise amplitude, which is likely due to the presence of air bubbles moving around the electrodes This effect is indeed more evident in dynamic conditions. Also, dynamic conditions introduce an aperiodic artefact, generating high-intensity peaks which cannot be attributed to the organoid physiological activity. As a consequence of such observations, the estimated values of the chosen parameters were likely overestimated. Despite that, consistent trends were obtained using both the planar and 3D multi-electrode array chip, with an almost constant burstiness index while the mean firing rate increases in dynamic with respect to static conditions. This is a coherent result, as the augmented oxygen supply due to advection is expected to enhance the firing frequency.
More rigorous strategies need to be developed for improving the signal-to-noise ratio using 3D multi-electrode arrays and to remove the aperiodic artefact, in order to obtain more reliable estimates of functional parameters. In addition, further acquisitions need to be performed, to perform statistical assessment of the obtained results with sufficient statistical power.
Although several limitations hold, this thesis could pave the way for advanced in vitro systems integrating multi-electrode arrays with fluidic platforms. These systems would improve the predictivity and translatability of electrophysiological characterization of brain organoids, allowing for recordings in conditions which more closely mimic those experienced by cells in vivo.